• Tweet this page
• Like this page
• Email this page
• Share this page

• Home»
• Publications »
• Books & monographs »
• Fauna of New Zealand Series »
• Extracts »
• FNZ68: Simuliidae »
• Molecular analysis of New Zealand Austrosimulium

FNZ 68 - Simuliidae (Insecta: Diptera) - Molecular analysis

Craig, DA, Cywinska A. 2012. Molecular analysis of New Zealand Austrosimulium (Diptera: Simuliiidae), pp. 60-65, 205-208, 298-301 in Craig DA, Craig REG, Crosby TK 2012. Simuliidae (Insecta: Diptera). Fauna of New Zealand 68, 336 pages.
( ISSN 0111-5383; no. 68 (print), ISSN 1179-7193 (online) ; no. 68. ISBN 978-0-478-34734-0 (print), ISBN 978-0-478-34735-7 (online) ). Published 29 June 2012
ZooBank: http://zoobank.org/References/9C478D54-FEB2-45E8-B61C-A3A06D4EB45D

Molecular analysis of New Zealand Austrosimulium (Diptera: Simuliidae) species

Douglas A. Craig
Department of Biological Sciences, University of Alberta, Edmonton, Alberta T6G 2E9, Canada
and
Alina Cywinska
3164 Candela Drive, Mississauga, Ontario L5A 2T8, Canada

Introduction
Molecular investigation of Simuliidae taxonomy has been sporadic. While interfamilial relationships of the Culicomorpha are well investigated (Pawlowski et al. 1996; Wiegmann et al. 2011), Moulton (1997, 2000, 2003) is the chief source for relationships within Simuliidae. Intrafamilial investigations are reviewed by Phayuhasena et al. (2010). For Austrosimulium there are only studies by Ballard (1994) and by Moulton, both restricted to Australian exemplars of the genus. There has been no such work on the New Zealand Austrosimulium fauna. Ballard’s study is of interest in that he used the rRNA 12S gene. That aligned with morphological differences between Austrosimulium (N.) pestilens Mackerras & Mackerras and A. (N.) bancrofti (Taylor), but could not resolve cytoforms known for A. bancrofti (Ballard & Bedo 1991). This result is similar in many ways to the situation reported by Conflitti et al. (2010) in which molecular and cytological evidence was taxonomically contradictory.

Other works on simuliids, too, have shown confusing results. Krueger & Hennings (2006) in a study of the Simulium damnosum Theobald complex (Africa) found striking inconsistencies in tree topologies depending on sequences used, and there was little correlation to host preference, behaviour, or other ecological parameters. On the other hand, Ilmonen et al. (2009) successfully used multiple characters from cytology, cytochrome c-oxidase subunit 1 (CO1) gene sequences, ecology, and morphology to clarify species status within the European Simulium vernum Macquart group. Similarly, multiple genes were used successfully, for the most part, at the species level for Thai simuliids (Phayuhasena et al. 2010).

Here we report on an extensive molecular dataset from the mitochondrial DNA CO1 gene (standard CO1-5’ barcodes and CO1-3’ markers) and that of mt16S rRNA gene, for New Zealand Austrosimulium. Concordance between backbone topology for species relationships derived from CO1 sequences and those from morphology is high (cf Fig. 505–507, 508–514). The results from this section are used elsewhere in this monograph for biogeographic purposes.

We note that we had no Australian Austrosimulium species in the dataset, and likewise 3 New Zealand species were not included, as material was not available for A. extendorum and A. fiordense (tillyardianum-subgroup) or A. campbellense (ungulatum-subgroup). However, it is unknown whether the absence of these species biased the results, given there was marked concatenation of available species within the tillyardianum subgroup (Fig 508, 508a).

Similarly there was a dearth of Australian Austrosimulium taxa in the morphological cladistic analysis section in this monograph (p. 53). Indeed, that segregate of the genus needs to be taxonomically revised, morphological descriptions brought up-to-date, and then exemplars included in molecular analyses.

Methods
DNA extractions for individual simuliids were obtained from tissue contained in 1–3 legs of adult specimens, or from the thoraxes of pupae and larvae, all fixed in at least 90% ethanol. Some 350 individuals were analysed from 165 samples. Voucher specimens are deposited in the New Zealand Arthropod Collection (NZAC), Landcare Research, Auckland, New Zealand.

For each individual, 30 µL of total DNA was extracted using the GeneEluteTM Mammalian Genomic DNA Miniprep Kit (Sigma-Aldrich Co., St. Louis, MO, 2003). Three pairs of primers were used to amplify the DNA extracts for standard DNA barcodes from the 5’ region of mtCO1 gene, non-standard DNA barcodes from the 3’ region of mtCO1, to cover the whole mtCO1 gene, and mt16S rRNA. A pair of universal primers, LCO1490 (5’-GGTCAACAAATCATAAAGATATTGG-3’) and HCO2198 (5’-TAAACTTCAGGGTGACCAAAAAATCA-3’) (Folmer et al. 1994), was used to amplify the standard CO1-5’ barcodes, i.e., ca 650 bp fragments at the 5’-terminus of the mitochondrial gene for cytochrome c oxidase subunit 1 (CO1; Folmer standardised region), which were trimmed to 618 bp. The CO1-3’primers, C1-J-2195 (5’-TTGATTTTTTGGTCATCCAGAAGT-3’) (Simons et al. 1994) and UEA10 –tRNA Leu gene (5’-TCCAATGCACTAATCTGCCATATTA-3’) (Lunt et al.1996) were used to amplify an 804 bp fragment in the remaining part of the CO1 gene. The universal mt16S primers, 16SA-L (5’-CGCCTGTTTATCAAAAACAT-3’) and 16SB-H (5’-CCGGTCT GAACTCAGATCACGT-3’) (Palumbi et al. 1991), were used to produce DNA templates in the mt 16S gene region; originally 550 bp long they were subsequently trimmed to 502 bp.

Each PCR cocktail contained 2.3 µL of 10×PCR buffer, pH 8.3 (10 mM of Tris-HCl, pH 8.3; and 50 mM of KCl; 0.01% NP-40),1.3 µL of mM MgCl2, 200 µM of each NTP, 1 unit Taq polymerase, 0.3 µM of each primer, 1–5 µL of DNA template and the remaining volume of ddH2O up to 25 µL. The PCR thermal regime consisted of one cycle of 1 min at 95oC; 35 cycles of 1 min at 94oC; 1 min at 55oC; and 1.5 min at 72oC, and a final cycle of 7 min at 72oC. All PCR products were subjected to dye terminator cycle sequencing reactions (30 cycles, 55oC annealing), and sequenced on ABI 3730 automated sequencers, using terminators with Big Dye v. 3.1, forward and reverse primers.

Electropherograms of obtained sequences were edited and aligned with SequencherTM v.4.5 (Gene Codes Corporation, Ann Arbor, MI, USA). Pairwise nucleotide sequence divergences were calculated using the Kimura 2-parameter model (Kimura 1980), while neighbour-joining (NJ) analysis (Saitou & Nei 1987) in MEGA™ 5.5 was used to examine relationships among taxa and determine bootstrap support for lineages. The tree was rooted using Simulium latipes (Meigen).

All sequences obtained in this study and their numbers are available on the Barcode of Life Data System <http://www.boldsystems.org/> under “New Zealand Simuliidae” in the “Completed Project” section. The sequence numbers are also available from Appendix 3 (p. 205), and on the website <fnz.landcareresearch.co.nz>.

Results and Discussion
Our molecular screening of Austrosimulium worked reasonably well for most of the analysed species: the target DNA was easily recovered from small amounts of insect tissue and aligned for all Austrosimulium species. Two groups of mtCO1 markers, the standard CO1-5’ barcodes and the non-standard barcode CO1-3’ sequences, contained no indels, their alignments were straightforward, and they lacked nonsense codons and pseudogenes. The third marker, mt16S, showed eight gaps per sequence.

Guanine-cytosine (GC) content provides a swift insight into mitochondrial genome biodiversity and nucleotide usage. Mitochondrial genomes show considerable variation in the GC contents (reflecting horizontal gene transfer or mutational bias) within phyla (Clare et al. 2008), a characteristic that can be utilised to measure the diversity among mitochondrial genomes in taxonomic classifications (Cywinska et al. 2010). The whole mitochondrial GC content can be estimated from GC variability in the standardised CO1-5’ barcodes as it is well correlated with that of full mitochondrial genomes (Clare et al. 2008).

In this study the standard CO1-5’ barcodes demonstrated A+T bias (average 64.3% for all codons), and was especially strong at third codon positions (average 82.4%) for all Austrosimulium species. The nucleotide composition varied slightly among species, with A+T content ranging in standard CO1-5’ barcodes from 63.1% (A. stewartense) to 65.4% (A. australense) for all codons, and from 79.5% to 85.2% at their third codon positions. Non-standard CO1-3’ barcodes also showed a significant A+T bias, with an average A+T content of 65.2% for all codons and 79.0% at the third codon positions.

Thus, the analysis of the Austrosimulium standard barcode CO1-5’ and non-standard CO1-3’ regions showed that the average GC content (35–36%) for all codon positions can be placed slightly above the centre of 22–45% GC range for the phylum Insecta (Clare et al. 2008). The average GC content for Austrosimulium species was somewhat larger than that observed in standard CO1-5’ barcodes for Tabanidae (32%; Cywinska et al. 2010), Canadian Culicidae (33%; Cywinska et al. 2006), and for Chironomidae (Ekrem et al. 2007).

The Austrosimulium CO1 sequences show skewed GC content distribution, with low GC presence at the third codon positions (average 18–21% for CO1-5’ and CO1-3’, respectively), which points to a strong shift in nucleotide usage at those sites. Still, the GC content in Austrosimulium is less skewed than that in mtCO1-5’ of Tabanidae with an average 3% GC at the third codon position (Cywinska et al. 2010).

Conspecific K2P divergence for Austrosimulium barcode CO1-5’ sequences averaged 2.3% (range 0–6.9%). Congeneric average divergence was at 6.7% (range 0.9%– 11.4%). For non-standard CO1-3’ barcodes, the mean conspecific divergence averaged 1.4% (range 0–14.2%) and congeneric divergence averaged 7.5% (range 0.8–13.6%). Thus, conspecific average values were lower for CO1-3’ sequences in comparison to CO1-5’ sequences, with wider range of intraspecific divergence.

Austrosimulium australense, represented here by 125 specimens, showed the highest interspecific similarity in the CO1-3’ region, with 1.9% average sequence divergence (range 0– 4.5%). For the same CO1-3’ region, A. longicorne (14 specimens) and A. ungulatum (38 individuals) showed less tight conspecific groupings, with relatively high average intraspecific sequence divergence of 3.6% (range 0–9.2%) and 2.6% (range 0–14.2%) respectively.

The average conspecific values were much higher for Austrosimulium species than, for example, for North American mosquitoes (0.5% K2P for CO1-5’ Cywinska et al. 2006), tabanids (0.49% for CO1-5’’ and 0.39% for CO1-3’’; Cywinska et al. 2010), North American birds (0.27%; Hebert et al. 2004), and moths (0.25%; Hebert et al. 2003).

Likewise, the non-standard CO1-3’ sequences for the outgroup species Simulium latipes used in this study showed divergences 5–7 times lower than for Austrosimulium with an average of 0.31% (range 0 – 6.3%).

The average congeneric divergences for Austrosimulium (6.7% K2P for CO1-5’ and 7.5% for CO1-3’) were similar to those for other simuliids (4% K2P for CO1-5’ and 6.5% for CO1-3’; Ilmonen et al. 2009), mosquitoes (10.4% K2P for CO1-5’; Cywinska et al. 2006), and tabanids (6% K2P for CO1-5’ and 9% for CO1-3’; Cywinska et al. 2010).

For the 16S gene, the average values for divergence within species and genus were 0.16% (range 0–0.8%) and 0.6% (range 0.1–1.9%), respectively, very low in comparison to CO1 sequences. As a result, NJ analysis of the 16S gene concatenated species and no comment on haplotypes can be made with confidence with the use of this gene. The 16S-based tree is not illustrated here, but is available online in Supplementary Data of this monograph; <fnz.landcareresearch.co.nz>.

Because of the size of the non-standard CO1-3’ NJ Tree (Fig 508, 509), we illustrate it in sections (Fig. 510–514) when showing phylogenetic relationships among haplotypes.

Given the generally assumed conservative evolutionary nature of the 16S ribosomal gene (Simon et al. 1994; Trewick & Wallis 2001) and its conservative nature in this study, we assume that grouping of some of the New Zealand Austrosimulium species together tends to indicate that species in the subgenus Novaustrosimulium (see Ballard 1994) indicates the latter is the older taxon —in agreement with more basal placement in the morphological cladistic analysis section of this monograph. Of relevance to that assertion, the paramere of male Novaustrosimulium species is moderately developed and with spines. We recognise and number some 43 haplotype lineages within both the australense and ungulatum species-groups, occasionally arbitrarily, and mainly to assist discussion of the results.

For the CO1 mitochondrial gene, rooted NJ analysis of CO1-3’ barcodes produced a tree with a backbone topology and bootstrap support which concurred well with those of the cladistic analysis of morphological characters used in this monograph (cf Fig 505, 508a, b & 509a, b). Of the three molecular markers tested, the non-standard barcode CO1-3’ locus was more effective as a diagnostic tool than the standard barcode CO1-5’, and much more effective than the more slowly evolving mt16S locus. The CO1-3’ locus, with its relatively wide gap between the interspecific and intraspecific sequence divergence, more frequent amino acid changes, comparatively high number of transversional and transitional substitutions, and slightly slower rates of transitional saturation, offers a little more flexibility in the interpretation of pairwise comparisons at the conspecific and congeneric levels than the standard barcode CO1-5’ locus (Cywinska et al. 2010).

Therefore, we focused exclusively on CO1-3’ barcodes in our molecular analysis of Austrosimulium taxonomy. In comparison to the cladistic morphological analysis section of this monograph there were two main differences:
1. The currently recognised unicorne-subgroup (ungulatum species-group) was sister to all others and included two species (A. dumbletoni, A. vailavoense) currently assignable on morphological grounds to the ungulatum-subgroup (Fig. 509a, 509b, 514).
2. The australense-subgroup with its two constituent species (A. australense, A. longicorne), usually considered as sister taxa, resolved A. australense as a pair of cryptic species with A. longicorne sister to the South Island segregate (Fig. 510).

Taxa of the tillyardianum-subgroup of species examined here (A. dugdalei, A. laticorne, A. multicorne, A. stewartense, A. tillyardianum) are not usefully aggregated by species or haplotype. So, apart from Fig. 508a, 508b, they are not further illustrated here in detail, but again, are available online in Supplementary Data for this monograph, <fnz.landcareresearch.co.nz>).

australense species-group
tillyardianum-subgroup (Fig. 508a, 508b)
This clade is 4.5% divergent from the australense-subgroup and with high support for monophyly (at 99%). Sister to the subgroup is an A. tillyardianum haplotype (NZS62, Pahau River Bridge, SH7, Canterbury) divergent from other haplotypes by ~6%. Perhaps this indicates a separate species, but material from that locality was of classic A. tillyardianum morphology. The sister clade, with high support, consisted of two poorly divergent lineages (~1%) with little better internal differentiation, or groupings. A major problem involves clustering together of morphologically markedly distinct species, for example, A. laticorne, A. tillyardianum, and A. multicorne (cf Fig. 275, 276, 279 of pupal gills). With such poor resolution at the species level, we feel any major comment on these species is not warranted even though this clade constitutes the majority of New Zealand simuliids. Hence, we do not provide further illustrations of the subgroup beyond Fig. 508, 508a. There are, however, three distinct clusters of A. tillyardianum, although for reasons stated above, we do not assign them much credence. One, for the South Island, comprised six samples (NZS2a, 12, 13, 63, 68, 84) and had a restricted distribution ranging from the Marlborough Sounds, south to Kaikoura and inland to Owen River Bridge (Buller River). Given the broad distribution of A. tillyardianum in the South Island (Map 14) this may represent the distribution of a distinct haplotype.

Two larger North Island clusters of A. tillyardianum (i.e., NZN7, 8, 10, 23, 36, 38, 42, 58, 59, 87, 91, and NZN3, 20, 34, 35, 37, 39, 40, 44, 62, 65, 67, 86, 88, 91) poorly divergent from one another are both equally well distributed over the full range of A. tillyardianum in the North Island. Indicative perhaps of recent origin, or, superior dispersal ability. Other minor clusterings of haplotypes (e.g., NZN11, 31, 41, 61) are similarly widespread across the full range of A. tillyardianum.

australense-subgroup (Fig. 510–513)
This well supported clade (A. australense + A. longicorne) is distinguished from the tillyardianum-subgroup with ca 4.5% divergence. The sister lineage, haplotype #1 (NZS103, Takaka Hill Walkway), with 1% divergence is somewhat problematic, since it was identified as A. longicorne. The material was, however, penultimate instar larvae, difficult to identify and may represent a cryptic species. The habitat was unusual. Otherwise, there is clear separation of North and South Island A. australense haplotypes into two, morphologically cryptic segregates, with 2% divergence, and again, strong support. Of significance is that no South Island haplotype occurred in the North Island, but there were four examples of North Island haplotypes occurring in the South Island (Fig. 509a, 509b, 511–513); none were very closely related to each other (discussed in more detail later). Of further note is that A. longicorne is sister to the South Island A. australense clade, but at only 1% divergence. That lower divergence is perhaps surprising given the major morphological difference between pupal gills of the two species (cf. Fig. 160, 161).

The A. longicorne clade is moderately well supported, indicating monophyly — in good agreement with morphology. There are two poorly supported sister lineages. One (haplotype #8, NZS14, 41), has the two constituent populations widely separated geographically. Notable is that NZS41 is from high altitude in the Old Man Range, Otago, while NZS14 is from close to sea level at Christchurch. The second lineage we consider to have three haplotypes (#9, 9a, 10); the first two from the South Island, divergent at 0.8 % from the well supported North Island haplotype (#10), which has little internal divergence.

Regarding the marked morphological divergence between pupal gills of A. australense and A. longicorne, it is now well established (Heming 2003; Carroll 2008) that a minor change in a regulator gene, can result in major differences in a structure. In the present instance (cf. Fig. 268, 269), this would be a reduction in the size of the pupal gill horn and a reduction in the number of filaments with increase in their thickness — nothing markedly difficult developmentally, but most different in final appearance. It would be phylogenetically useful to examine the ontogeny of the pupal gills of these two species.

Lineages of A. australense in the South Island have little support. The sister haplotype (#7) is from Kawhaka Creek, West Coast (NZS50). Its sister lineage we consider as five haplotypes (#2–6). Haplotype #2 (NZS51, 67, 70, 72, 73, 74, 81, 99) is restricted to the northwest of the island, ranging from the Farewell Spit region, south to Greymouth, and inland to the Rahu Saddle. The next, #3 (NZS 1, 15, 16, 54, 77, 86,101), is more widespread, from northwest Nelson to Marlborough and south to the Canterbury Plains, and is likely to be the haplotype of the synonymic name tillyardi Tonnoir, 1923 from Nelson. The haplotype sister to these two groups is #4 from Inangahua Junction (NZS68). The remaining major grouping is of two poorly discriminated lineages of haplotypes #5 (NZS79, 102) and #6 (NZS2, 12, 78, 82, 89) which are widespread in the north of the island, but extend south only to Green Burn, Kaikoura (NZS12).

The well supported North Island clade of A. australense has a number of haplotypes, most poorly discriminated (Fig. 511–513). Some have, however, distinct distributions. We consider the first three lineages as one haplotype viz. #11, consisting of NZN41 (Ohiwa Stream, Hastings-Taihape Road) and two from NZN63 (near Tikitiki, SH35) — an eastern distribution. One clade of haplotype #12 (NZN 70, 72, 76, 77, 82, 94) and #13 (NZN9, 84a) is largely found north from Rotorua into Northland. Any of haplotypes #12, 28, or 32 could be the haplotype for the original material named australense by Schiner (1868) — all are found in the Waitakere Ranges (NZN93, 94; Fig. 511, 513). There is one southern outlier of #13 at NZN84a (Mount Ruapehu). This sample is problematic in that morphologically, it is definitively A. longicorne. The habitat is not typical for A. australense either (similar to Fig. 465). Because this population occurs well south of others of that haplotype, this may be an error in analysis.

Sister to the remaining haplotypes is that from Stewart Island (#14 NZS170, Kaipipi Inlet). Something of a conundrum, this datum may well be correct given that North Island haplotype #24 (NZS29) occurs in the Catlins, southern South Island (Fig 512). Another lineage #16 (NZN3, 7, 15, 32a, 53, 61, 62, 92) has a distribution that is similar to other haplotypes, i.e., a peripheral distribution around the central portion of the North Island. The distribution of this lineage is the base of Coromandel Peninsula, East Cape, Hawkes Bay, Taranaki, and Raglan. One population from Ohakune (NZN92), more centrally on the Volcanic Plateau, is anomalous, but it is not a misidentification. A small clade of five haplotypes #17 (NZN 4, 8, 65) and its sister, #17a of two haplotypes (NZN57, 78) has a more restricted northern distribution ranging from far Northland, Coromandel Peninsula, Bay of Plenty, and East Cape. Sister to those two, #18 (NZN36, 64, 85), is distributed along the east coast from East Cape to Waihi Stream, SH52, Manawatu. One population (NZ85) is more central, just north of Taihape. Haplotype #19 (NZN63, 65, 73, 79, 80, 81, 83 ), that has markedly low internal divergence, has a distribution similar to haplotypes #17, 17a, and 18, occurring well north in Northland, south to the Bay of Islands, but with a gap through the Coromandel and Bay of Plenty to East Cape.

The specimens of the unavailable name ‘caecutiens’ Walker, 1848 (Fig. 519–521) would be part of haplotype #19. If future investigations show haplotype #19 to be taxonomically distinct from A. australense, then the ‘caecutiens’ specimens should be listed under the name assigned to haplotype #19.

Not markedly divergent from those haplotypes is that of #20, but this is a northwestern South Island population (NZS99, Green Hills, Farewell Spit region). Perhaps, again, this is a conundrum, but given a probable connection between the North and South Islands during glacial sea level depressions (Fig. 514, 516), its occurrence in the Farewell Spit area is not unexpected (and see below). That is, if simuliids manage to cross Cook Strait from the North Island, this is one area where it might have occurred.

A more southeastern haplotype in the North Island is #21 (NZS37, 39, 51), which ranges only from Hawkes Bay to near Dannevirke.

Sister to the remainder of the haplotypes is #22 (NZN30, Rangitikei River), a more south-central locality (Fig. 512). Haplotypes #23–25 show little divergence. That of #23 (NZN 20, 32, 52, 56, 58, 67, 75, 86, 89), occurs in Northland, Bay of Plenty, Hawkes Bay, Taranaki Bight, and Raglan, and again has a largely coastal distribution around the North Island. That haplotype is unresolved from what we call #25 (NZN54, 61, 74), which is found in Northland and southeastern East Cape. Probably haplotypes #23 and 25 should be considered as one. With low divergence from those haplotypes is #24 (NZN33, 35, and NZS29). Of note, of course, is that the latter population is from the Catlins, South Island.

Another tightly nested haplotype is #27, that again shows a largely peripheral distribution (Fig. 513) (NZN5, 6, 10, 11, 17, 21, 22, 24, 34, 40, 52, 71, 78, 90). Its range includes Northland, Coromandel, Hawkes Bay, Taranaki and two localities in southern Waikato. Haplotype #28 (NZN 19, 25, 27, 29, 43, 56, 93), has a more central, but still peripheral distribution that ranges from Auckland to Bay of Plenty, inland Hawkes Bay, Taranaki, and Raglan. There are no deep central localities.

A slightly more divergent group is #29 (NZN13, Raglan, NZN26, north of Mount Taranaki). Of biogeographic importance are haplotypes #30–32. Haplotype #31 (sister to #30), occurs at Totaranui, Golden Bay, South Island (NZS102); derived from probable connections between the two islands? An assumption might be that its two sister populations would be concentrated in the southern part of the North Island, but that is not so, and in fact they are in Northland (NZN75). The remaining haplotype #32 (NZN 17, 66, 69, 70, 76, 86, 93), while also with a peripheral distribution, is widespread, ranging from Northland to inland Bay of Plenty, East Cape and Raglan and, following a major gap, near Wellington.

ungulatum species-group (Fig. 514)
As noted elsewhere, the cladistic analysis of morphological characters section in this monograph showed the ungulatum species-group comprised two sister clades, the ungulatum- and unicorne- subgroups (Fig. 506). The molecular analysis is at slight variance to this.

unicorne-subgroup
This is the very well supported sister clade to all other New Zealand Austrosimulium, and is well discriminated at ca the 6% level from the ungulatum-subgroup and is not sister to it. The unicorne-subgroup here consists of A. vailavoense + A. unicorne as a clade with marked internal divergence (7.2%), sister to A. bicorne + A. dumbletoni + A. tonnoiri, also with moderate internal divergence. Austrosimulium vailavoense (haplotype #39) shows minor divergence between the population at Papatotara, Southland (NZS157) and the type locality at Vaila Voe Bay, Stewart Island (NZS165). In the sister clade, one population (haplotype #41) of A. bicorne at Temple Basin, Arthurs Pass (NZS133) is sister to the remainder of the clade at the 1% level. Another haplotype (#42) of A. bicorne from the Homer Tunnel (NZS32) shows minor divergence from its sister haplotype #43. That latter haplotype is shared among 3 species. This is not surprising given morphological similarities between A. tonnoiri and A. bicorne. Whether A. dumbletoni truly shares that haplotype is moot, but discovery of its immature stages and their habitat requirements would confirm, or not, that placement. Probable requirements of larvae and pupae of A. dumbletoni are discussed elsewhere in this monograph (p. 149).

ungulatum-subgroup
Sister to the australense species-group is a clade comprised of A. ungulatum + A. vexans. This is well separated from the australense species-group (tillyardianum- + australense-subgroups) at 5% and is fully supported. Two clades that diverge at the 1% level, however, are only moderately well supported. One has A. vexans sister (#34) to five lineages (NZS30, 46, 67, 157) of A. ungulatum, two from one locality. These are considered haplotype #33, and are currently all identified morphologically as classic A. ungulatum. If this is the correct arrangement, we consider those haplotypes to be a cryptic segregate of A. ungulatum that ranges from Arthurs Pass to mid-Westland and down to southern Southland.

The second, moderately well supported, and sister to all remaining A. ungulatum haplotypes, is haplotype #35, from near Greymouth (NZS82). Its sister clade is not well supported. We consider that clade to have three internal haplotypes, viz #s 36, 37, 38, all moderately-to-well supported. Haplotype #36 (NZS1, 27, 28, 35, 45, 45a, 65, 72, 79, 82, 83, 86, 104), ranges from Nelson to the west, down Westland to Jackson Bay, and in western and eastern Southland. A major part of that grouping is homogeneous. Haplotype #37 has but two widely separated populations at Greymouth (NZS33) and Hollyford Valley (NZS33). The last haplotype, #38 (NZS4, 34, 37, 47, 49, 54, 55, 56, 65, 66, 68, 70, 74, 75, 89, 106, 170), is in large part a grade. It ranges mainly over the whole of the South Island, even Stewart Island, and overlaps well with haplotype #36. As discussed elsewhere in this monograph, occurrence of that haplotype at Arthurs Pass, well indicates that a gap in its distribution on the West Coast is not the result of glaciation.

Purchase this publication

• Purchasing details